| 1 | /* handy.h |
| 2 | * |
| 3 | * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1999, 2000, |
| 4 | * 2001, 2002, 2004, 2005, 2006, 2007, 2008 by Larry Wall and others |
| 5 | * |
| 6 | * You may distribute under the terms of either the GNU General Public |
| 7 | * License or the Artistic License, as specified in the README file. |
| 8 | * |
| 9 | */ |
| 10 | |
| 11 | #if !defined(__STDC__) |
| 12 | #ifdef NULL |
| 13 | #undef NULL |
| 14 | #endif |
| 15 | #ifndef I286 |
| 16 | # define NULL 0 |
| 17 | #else |
| 18 | # define NULL 0L |
| 19 | #endif |
| 20 | #endif |
| 21 | |
| 22 | #ifndef PERL_CORE |
| 23 | # define Null(type) ((type)NULL) |
| 24 | |
| 25 | /* |
| 26 | =head1 Handy Values |
| 27 | |
| 28 | =for apidoc AmU||Nullch |
| 29 | Null character pointer. (No longer available when C<PERL_CORE> is defined.) |
| 30 | |
| 31 | =for apidoc AmU||Nullsv |
| 32 | Null SV pointer. (No longer available when C<PERL_CORE> is defined.) |
| 33 | |
| 34 | =cut |
| 35 | */ |
| 36 | |
| 37 | # define Nullch Null(char*) |
| 38 | # define Nullfp Null(PerlIO*) |
| 39 | # define Nullsv Null(SV*) |
| 40 | #endif |
| 41 | |
| 42 | #ifdef TRUE |
| 43 | #undef TRUE |
| 44 | #endif |
| 45 | #ifdef FALSE |
| 46 | #undef FALSE |
| 47 | #endif |
| 48 | #define TRUE (1) |
| 49 | #define FALSE (0) |
| 50 | |
| 51 | /* The MUTABLE_*() macros cast pointers to the types shown, in such a way |
| 52 | * (compiler permitting) that casting away const-ness will give a warning; |
| 53 | * e.g.: |
| 54 | * |
| 55 | * const SV *sv = ...; |
| 56 | * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away |
| 57 | * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn |
| 58 | */ |
| 59 | |
| 60 | #if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN) |
| 61 | # define MUTABLE_PTR(p) ({ void *_p = (p); _p; }) |
| 62 | #else |
| 63 | # define MUTABLE_PTR(p) ((void *) (p)) |
| 64 | #endif |
| 65 | |
| 66 | #define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p)) |
| 67 | #define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p)) |
| 68 | #define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p)) |
| 69 | #define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p)) |
| 70 | #define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p)) |
| 71 | #define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p)) |
| 72 | |
| 73 | #ifdef I_STDBOOL |
| 74 | # include <stdbool.h> |
| 75 | # ifndef HAS_BOOL |
| 76 | # define HAS_BOOL 1 |
| 77 | # endif |
| 78 | #endif |
| 79 | |
| 80 | /* bool is built-in for g++-2.6.3 and later, which might be used |
| 81 | for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't |
| 82 | be sure _G_config.h will be included before this file. _G_config.h |
| 83 | also defines _G_HAVE_BOOL for both gcc and g++, but only g++ |
| 84 | actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us. |
| 85 | g++ can be identified by __GNUG__. |
| 86 | Andy Dougherty February 2000 |
| 87 | */ |
| 88 | #ifdef __GNUG__ /* GNU g++ has bool built-in */ |
| 89 | # ifndef HAS_BOOL |
| 90 | # define HAS_BOOL 1 |
| 91 | # endif |
| 92 | #endif |
| 93 | |
| 94 | /* The NeXT dynamic loader headers will not build with the bool macro |
| 95 | So declare them now to clear confusion. |
| 96 | */ |
| 97 | #if defined(NeXT) || defined(__NeXT__) |
| 98 | # undef FALSE |
| 99 | # undef TRUE |
| 100 | typedef enum bool { FALSE = 0, TRUE = 1 } bool; |
| 101 | # define ENUM_BOOL 1 |
| 102 | # ifndef HAS_BOOL |
| 103 | # define HAS_BOOL 1 |
| 104 | # endif /* !HAS_BOOL */ |
| 105 | #endif /* NeXT || __NeXT__ */ |
| 106 | |
| 107 | #ifndef HAS_BOOL |
| 108 | # if defined(UTS) || defined(VMS) |
| 109 | # define bool int |
| 110 | # else |
| 111 | # define bool char |
| 112 | # endif |
| 113 | # define HAS_BOOL 1 |
| 114 | #endif |
| 115 | |
| 116 | /* a simple (bool) cast may not do the right thing: if bool is defined |
| 117 | * as char for example, then the cast from int is implementation-defined |
| 118 | * (bool)!!(cbool) in a ternary triggers a bug in xlc on AIX |
| 119 | */ |
| 120 | |
| 121 | #define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0) |
| 122 | |
| 123 | /* Try to figure out __func__ or __FUNCTION__ equivalent, if any. |
| 124 | * XXX Should really be a Configure probe, with HAS__FUNCTION__ |
| 125 | * and FUNCTION__ as results. |
| 126 | * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */ |
| 127 | #if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */ |
| 128 | # define FUNCTION__ __func__ |
| 129 | #else |
| 130 | # if (defined(_MSC_VER) && _MSC_VER < 1300) || /* Pre-MSVC 7.0 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \ |
| 131 | (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */ |
| 132 | # define FUNCTION__ "" |
| 133 | # else |
| 134 | # define FUNCTION__ __FUNCTION__ /* Common extension. */ |
| 135 | # endif |
| 136 | #endif |
| 137 | |
| 138 | /* XXX A note on the perl source internal type system. The |
| 139 | original intent was that I32 be *exactly* 32 bits. |
| 140 | |
| 141 | Currently, we only guarantee that I32 is *at least* 32 bits. |
| 142 | Specifically, if int is 64 bits, then so is I32. (This is the case |
| 143 | for the Cray.) This has the advantage of meshing nicely with |
| 144 | standard library calls (where we pass an I32 and the library is |
| 145 | expecting an int), but the disadvantage that an I32 is not 32 bits. |
| 146 | Andy Dougherty August 1996 |
| 147 | |
| 148 | There is no guarantee that there is *any* integral type with |
| 149 | exactly 32 bits. It is perfectly legal for a system to have |
| 150 | sizeof(short) == sizeof(int) == sizeof(long) == 8. |
| 151 | |
| 152 | Similarly, there is no guarantee that I16 and U16 have exactly 16 |
| 153 | bits. |
| 154 | |
| 155 | For dealing with issues that may arise from various 32/64-bit |
| 156 | systems, we will ask Configure to check out |
| 157 | |
| 158 | SHORTSIZE == sizeof(short) |
| 159 | INTSIZE == sizeof(int) |
| 160 | LONGSIZE == sizeof(long) |
| 161 | LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG) |
| 162 | PTRSIZE == sizeof(void *) |
| 163 | DOUBLESIZE == sizeof(double) |
| 164 | LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE). |
| 165 | |
| 166 | */ |
| 167 | |
| 168 | #ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */ |
| 169 | # include <inttypes.h> |
| 170 | # ifdef INT32_MIN_BROKEN |
| 171 | # undef INT32_MIN |
| 172 | # define INT32_MIN (-2147483647-1) |
| 173 | # endif |
| 174 | # ifdef INT64_MIN_BROKEN |
| 175 | # undef INT64_MIN |
| 176 | # define INT64_MIN (-9223372036854775807LL-1) |
| 177 | # endif |
| 178 | #endif |
| 179 | |
| 180 | typedef I8TYPE I8; |
| 181 | typedef U8TYPE U8; |
| 182 | typedef I16TYPE I16; |
| 183 | typedef U16TYPE U16; |
| 184 | typedef I32TYPE I32; |
| 185 | typedef U32TYPE U32; |
| 186 | #ifdef PERL_CORE |
| 187 | # ifdef HAS_QUAD |
| 188 | typedef I64TYPE I64; |
| 189 | typedef U64TYPE U64; |
| 190 | # endif |
| 191 | #endif /* PERL_CORE */ |
| 192 | |
| 193 | #if defined(HAS_QUAD) && defined(USE_64_BIT_INT) |
| 194 | # ifndef UINT64_C /* usually from <inttypes.h> */ |
| 195 | # if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG |
| 196 | # define INT64_C(c) CAT2(c,LL) |
| 197 | # define UINT64_C(c) CAT2(c,ULL) |
| 198 | # else |
| 199 | # if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG |
| 200 | # define INT64_C(c) CAT2(c,L) |
| 201 | # define UINT64_C(c) CAT2(c,UL) |
| 202 | # else |
| 203 | # if defined(_WIN64) && defined(_MSC_VER) |
| 204 | # define INT64_C(c) CAT2(c,I64) |
| 205 | # define UINT64_C(c) CAT2(c,UI64) |
| 206 | # else |
| 207 | # define INT64_C(c) ((I64TYPE)(c)) |
| 208 | # define UINT64_C(c) ((U64TYPE)(c)) |
| 209 | # endif |
| 210 | # endif |
| 211 | # endif |
| 212 | # endif |
| 213 | #endif |
| 214 | |
| 215 | #if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX) |
| 216 | |
| 217 | /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type. |
| 218 | Please search CHAR_MAX in perl.h for further details. */ |
| 219 | #define U8_MAX UINT8_MAX |
| 220 | #define U8_MIN UINT8_MIN |
| 221 | |
| 222 | #define I16_MAX INT16_MAX |
| 223 | #define I16_MIN INT16_MIN |
| 224 | #define U16_MAX UINT16_MAX |
| 225 | #define U16_MIN UINT16_MIN |
| 226 | |
| 227 | #define I32_MAX INT32_MAX |
| 228 | #define I32_MIN INT32_MIN |
| 229 | #ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */ |
| 230 | # define U32_MAX UINT32_MAX |
| 231 | #else |
| 232 | # define U32_MAX 4294967295U |
| 233 | #endif |
| 234 | #define U32_MIN UINT32_MIN |
| 235 | |
| 236 | #else |
| 237 | |
| 238 | /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type. |
| 239 | Please search CHAR_MAX in perl.h for further details. */ |
| 240 | #define U8_MAX PERL_UCHAR_MAX |
| 241 | #define U8_MIN PERL_UCHAR_MIN |
| 242 | |
| 243 | #define I16_MAX PERL_SHORT_MAX |
| 244 | #define I16_MIN PERL_SHORT_MIN |
| 245 | #define U16_MAX PERL_USHORT_MAX |
| 246 | #define U16_MIN PERL_USHORT_MIN |
| 247 | |
| 248 | #if LONGSIZE > 4 |
| 249 | # define I32_MAX PERL_INT_MAX |
| 250 | # define I32_MIN PERL_INT_MIN |
| 251 | # define U32_MAX PERL_UINT_MAX |
| 252 | # define U32_MIN PERL_UINT_MIN |
| 253 | #else |
| 254 | # define I32_MAX PERL_LONG_MAX |
| 255 | # define I32_MIN PERL_LONG_MIN |
| 256 | # define U32_MAX PERL_ULONG_MAX |
| 257 | # define U32_MIN PERL_ULONG_MIN |
| 258 | #endif |
| 259 | |
| 260 | #endif |
| 261 | |
| 262 | /* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */ |
| 263 | #define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */ |
| 264 | #define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8) |
| 265 | #define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */ |
| 266 | |
| 267 | #define Ctl(ch) ((ch) & 037) |
| 268 | |
| 269 | /* |
| 270 | =head1 SV-Body Allocation |
| 271 | |
| 272 | =for apidoc Ama|SV*|newSVpvs|const char* s |
| 273 | Like C<newSVpvn>, but takes a literal string instead of a string/length pair. |
| 274 | |
| 275 | =for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags |
| 276 | Like C<newSVpvn_flags>, but takes a literal string instead of a string/length |
| 277 | pair. |
| 278 | |
| 279 | =for apidoc Ama|SV*|newSVpvs_share|const char* s |
| 280 | Like C<newSVpvn_share>, but takes a literal string instead of a string/length |
| 281 | pair and omits the hash parameter. |
| 282 | |
| 283 | =for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags |
| 284 | Like C<sv_catpvn_flags>, but takes a literal string instead of a |
| 285 | string/length pair. |
| 286 | |
| 287 | =for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s |
| 288 | Like C<sv_catpvn_nomg>, but takes a literal string instead of a |
| 289 | string/length pair. |
| 290 | |
| 291 | =for apidoc Am|void|sv_catpvs|SV* sv|const char* s |
| 292 | Like C<sv_catpvn>, but takes a literal string instead of a string/length pair. |
| 293 | |
| 294 | =for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s |
| 295 | Like C<sv_catpvn_mg>, but takes a literal string instead of a |
| 296 | string/length pair. |
| 297 | |
| 298 | =for apidoc Am|void|sv_setpvs|SV* sv|const char* s |
| 299 | Like C<sv_setpvn>, but takes a literal string instead of a string/length pair. |
| 300 | |
| 301 | =for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s |
| 302 | Like C<sv_setpvn_mg>, but takes a literal string instead of a |
| 303 | string/length pair. |
| 304 | |
| 305 | =for apidoc Am|SV *|sv_setref_pvs|const char* s |
| 306 | Like C<sv_setref_pvn>, but takes a literal string instead of a |
| 307 | string/length pair. |
| 308 | |
| 309 | =head1 Memory Management |
| 310 | |
| 311 | =for apidoc Ama|char*|savepvs|const char* s |
| 312 | Like C<savepvn>, but takes a literal string instead of a string/length pair. |
| 313 | |
| 314 | =for apidoc Ama|char*|savesharedpvs|const char* s |
| 315 | A version of C<savepvs()> which allocates the duplicate string in memory |
| 316 | which is shared between threads. |
| 317 | |
| 318 | =head1 GV Functions |
| 319 | |
| 320 | =for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create |
| 321 | Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair. |
| 322 | |
| 323 | =head1 Hash Manipulation Functions |
| 324 | |
| 325 | =for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval |
| 326 | Like C<hv_fetch>, but takes a literal string instead of a string/length pair. |
| 327 | |
| 328 | =for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val |
| 329 | Like C<hv_store>, but takes a literal string instead of a string/length pair |
| 330 | and omits the hash parameter. |
| 331 | |
| 332 | =head1 Lexer interface |
| 333 | |
| 334 | =for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags |
| 335 | |
| 336 | Like L</lex_stuff_pvn>, but takes a literal string instead of a |
| 337 | string/length pair. |
| 338 | |
| 339 | =cut |
| 340 | */ |
| 341 | |
| 342 | /* concatenating with "" ensures that only literal strings are accepted as argument */ |
| 343 | #define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1) |
| 344 | |
| 345 | /* note that STR_WITH_LEN() can't be used as argument to macros or functions that |
| 346 | * under some configurations might be macros, which means that it requires the full |
| 347 | * Perl_xxx(aTHX_ ...) form for any API calls where it's used. |
| 348 | */ |
| 349 | |
| 350 | /* STR_WITH_LEN() shortcuts */ |
| 351 | #define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str)) |
| 352 | #define newSVpvs_flags(str,flags) \ |
| 353 | Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags) |
| 354 | #define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0) |
| 355 | #define sv_catpvs_flags(sv, str, flags) \ |
| 356 | Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags) |
| 357 | #define sv_catpvs_nomg(sv, str) \ |
| 358 | Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0) |
| 359 | #define sv_catpvs(sv, str) \ |
| 360 | Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC) |
| 361 | #define sv_catpvs_mg(sv, str) \ |
| 362 | Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC) |
| 363 | #define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str)) |
| 364 | #define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str)) |
| 365 | #define sv_setref_pvs(rv, classname, str) \ |
| 366 | Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str)) |
| 367 | #define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str)) |
| 368 | #define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str)) |
| 369 | #define gv_stashpvs(str, create) \ |
| 370 | Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create) |
| 371 | #define gv_fetchpvs(namebeg, add, sv_type) \ |
| 372 | Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type) |
| 373 | #define gv_fetchpvn(namebeg, len, add, sv_type) \ |
| 374 | Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type) |
| 375 | #define sv_catxmlpvs(dsv, str, utf8) \ |
| 376 | Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8) |
| 377 | #define hv_fetchs(hv,key,lval) \ |
| 378 | ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \ |
| 379 | (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \ |
| 380 | : HV_FETCH_JUST_SV, NULL, 0)) |
| 381 | |
| 382 | #define hv_stores(hv,key,val) \ |
| 383 | ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \ |
| 384 | (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0)) |
| 385 | |
| 386 | #define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags) |
| 387 | |
| 388 | #define get_cvs(str, flags) \ |
| 389 | Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags)) |
| 390 | |
| 391 | /* |
| 392 | =head1 Miscellaneous Functions |
| 393 | |
| 394 | =for apidoc Am|bool|strNE|char* s1|char* s2 |
| 395 | Test two strings to see if they are different. Returns true or |
| 396 | false. |
| 397 | |
| 398 | =for apidoc Am|bool|strEQ|char* s1|char* s2 |
| 399 | Test two strings to see if they are equal. Returns true or false. |
| 400 | |
| 401 | =for apidoc Am|bool|strLT|char* s1|char* s2 |
| 402 | Test two strings to see if the first, C<s1>, is less than the second, |
| 403 | C<s2>. Returns true or false. |
| 404 | |
| 405 | =for apidoc Am|bool|strLE|char* s1|char* s2 |
| 406 | Test two strings to see if the first, C<s1>, is less than or equal to the |
| 407 | second, C<s2>. Returns true or false. |
| 408 | |
| 409 | =for apidoc Am|bool|strGT|char* s1|char* s2 |
| 410 | Test two strings to see if the first, C<s1>, is greater than the second, |
| 411 | C<s2>. Returns true or false. |
| 412 | |
| 413 | =for apidoc Am|bool|strGE|char* s1|char* s2 |
| 414 | Test two strings to see if the first, C<s1>, is greater than or equal to |
| 415 | the second, C<s2>. Returns true or false. |
| 416 | |
| 417 | =for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len |
| 418 | Test two strings to see if they are different. The C<len> parameter |
| 419 | indicates the number of bytes to compare. Returns true or false. (A |
| 420 | wrapper for C<strncmp>). |
| 421 | |
| 422 | =for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len |
| 423 | Test two strings to see if they are equal. The C<len> parameter indicates |
| 424 | the number of bytes to compare. Returns true or false. (A wrapper for |
| 425 | C<strncmp>). |
| 426 | |
| 427 | =cut |
| 428 | */ |
| 429 | |
| 430 | #define strNE(s1,s2) (strcmp(s1,s2)) |
| 431 | #define strEQ(s1,s2) (!strcmp(s1,s2)) |
| 432 | #define strLT(s1,s2) (strcmp(s1,s2) < 0) |
| 433 | #define strLE(s1,s2) (strcmp(s1,s2) <= 0) |
| 434 | #define strGT(s1,s2) (strcmp(s1,s2) > 0) |
| 435 | #define strGE(s1,s2) (strcmp(s1,s2) >= 0) |
| 436 | #define strnNE(s1,s2,l) (strncmp(s1,s2,l)) |
| 437 | #define strnEQ(s1,s2,l) (!strncmp(s1,s2,l)) |
| 438 | |
| 439 | #ifdef HAS_MEMCMP |
| 440 | # define memNE(s1,s2,l) (memcmp(s1,s2,l)) |
| 441 | # define memEQ(s1,s2,l) (!memcmp(s1,s2,l)) |
| 442 | #else |
| 443 | # define memNE(s1,s2,l) (bcmp(s1,s2,l)) |
| 444 | # define memEQ(s1,s2,l) (!bcmp(s1,s2,l)) |
| 445 | #endif |
| 446 | |
| 447 | #define memEQs(s1, l, s2) \ |
| 448 | (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1))) |
| 449 | #define memNEs(s1, l, s2) !memEQs(s1, l, s2) |
| 450 | |
| 451 | /* |
| 452 | * Character classes. |
| 453 | * |
| 454 | * Unfortunately, the introduction of locales means that we |
| 455 | * can't trust isupper(), etc. to tell the truth. And when |
| 456 | * it comes to /\w+/ with tainting enabled, we *must* be able |
| 457 | * to trust our character classes. |
| 458 | * |
| 459 | * Therefore, the default tests in the text of Perl will be |
| 460 | * independent of locale. Any code that wants to depend on |
| 461 | * the current locale will use the tests that begin with "lc". |
| 462 | */ |
| 463 | |
| 464 | #ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */ |
| 465 | # ifndef CTYPE256 |
| 466 | # define CTYPE256 |
| 467 | # endif |
| 468 | #endif |
| 469 | |
| 470 | /* |
| 471 | |
| 472 | =head1 Character classes |
| 473 | There are three variants for all the functions in this section. The base ones |
| 474 | operate using the character set of the platform Perl is running on. The ones |
| 475 | with an C<_A> suffix operate on the ASCII character set, and the ones with an |
| 476 | C<_L1> suffix operate on the full Latin1 character set. All are unaffected by |
| 477 | locale and by C<use bytes>. |
| 478 | |
| 479 | For ASCII platforms, the base function with no suffix and the one with the |
| 480 | C<_A> suffix are identical. The function with the C<_L1> suffix imposes the |
| 481 | Latin-1 character set onto the platform. That is, the code points that are |
| 482 | ASCII are unaffected, since ASCII is a subset of Latin-1. But the non-ASCII |
| 483 | code points are treated as if they are Latin-1 characters. For example, |
| 484 | C<isSPACE_L1()> will return true when called with the code point 0xA0, which is |
| 485 | the Latin-1 NO-BREAK SPACE. |
| 486 | |
| 487 | For EBCDIC platforms, the base function with no suffix and the one with the |
| 488 | C<_L1> suffix should be identical, since, as of this writing, the EBCDIC code |
| 489 | pages that Perl knows about all are equivalent to Latin-1. The function that |
| 490 | ends in an C<_A> suffix will not return true unless the specified character also |
| 491 | has an ASCII equivalent. |
| 492 | |
| 493 | =for apidoc Am|bool|isALPHA|char ch |
| 494 | Returns a boolean indicating whether the specified character is an |
| 495 | alphabetic character in the platform's native character set. |
| 496 | See the L<top of this section|/Character classes> for an explanation of variants |
| 497 | C<isALPHA_A> and C<isALPHA_L1>. |
| 498 | |
| 499 | =for apidoc Am|bool|isASCII|char ch |
| 500 | Returns a boolean indicating whether the specified character is one of the 128 |
| 501 | characters in the ASCII character set. On non-ASCII platforms, it is if this |
| 502 | character corresponds to an ASCII character. Variants C<isASCII_A()> and |
| 503 | C<isASCII_L1()> are identical to C<isASCII()>. |
| 504 | |
| 505 | =for apidoc Am|bool|isDIGIT|char ch |
| 506 | Returns a boolean indicating whether the specified character is a |
| 507 | digit in the platform's native character set. |
| 508 | Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>. |
| 509 | |
| 510 | =for apidoc Am|bool|isLOWER|char ch |
| 511 | Returns a boolean indicating whether the specified character is a |
| 512 | lowercase character in the platform's native character set. |
| 513 | See the L<top of this section|/Character classes> for an explanation of variants |
| 514 | C<isLOWER_A> and C<isLOWER_L1>. |
| 515 | |
| 516 | =for apidoc Am|bool|isOCTAL|char ch |
| 517 | Returns a boolean indicating whether the specified character is an |
| 518 | octal digit, [0-7] in the platform's native character set. |
| 519 | Variants C<isOCTAL_A> and C<isOCTAL_L1> are identical to C<isOCTAL>. |
| 520 | |
| 521 | =for apidoc Am|bool|isSPACE|char ch |
| 522 | Returns a boolean indicating whether the specified character is a |
| 523 | whitespace character in the platform's native character set. This is the same |
| 524 | as what C<\s> matches in a regular expression. |
| 525 | See the L<top of this section|/Character classes> for an explanation of variants |
| 526 | C<isSPACE_A> and C<isSPACE_L1>. |
| 527 | |
| 528 | =for apidoc Am|bool|isUPPER|char ch |
| 529 | Returns a boolean indicating whether the specified character is an |
| 530 | uppercase character in the platform's native character set. |
| 531 | See the L<top of this section|/Character classes> for an explanation of variants |
| 532 | C<isUPPER_A> and C<isUPPER_L1>. |
| 533 | |
| 534 | =for apidoc Am|bool|isWORDCHAR|char ch |
| 535 | Returns a boolean indicating whether the specified character is a |
| 536 | character that is any of: alphabetic, numeric, or an underscore. This is the |
| 537 | same as what C<\w> matches in a regular expression. |
| 538 | C<isALNUM()> is a synonym provided for backward compatibility. Note that it |
| 539 | does not have the standard C language meaning of alphanumeric, since it matches |
| 540 | an underscore and the standard meaning does not. |
| 541 | See the L<top of this section|/Character classes> for an explanation of variants |
| 542 | C<isWORDCHAR_A> and C<isWORDCHAR_L1>. |
| 543 | |
| 544 | =for apidoc Am|bool|isXDIGIT|char ch |
| 545 | Returns a boolean indicating whether the specified character is a hexadecimal |
| 546 | digit, [0-9A-Fa-f]. Variants C<isXDIGIT_A()> and C<isXDIGIT_L1()> are |
| 547 | identical to C<isXDIGIT()>. |
| 548 | |
| 549 | =head1 Character case changing |
| 550 | |
| 551 | =for apidoc Am|char|toUPPER|char ch |
| 552 | Converts the specified character to uppercase in the platform's native |
| 553 | character set, if possible; otherwise returns the input character itself. |
| 554 | |
| 555 | =for apidoc Am|char|toLOWER|char ch |
| 556 | Converts the specified character to lowercase in the platform's native |
| 557 | character set, if possible; otherwise returns the input character itself. |
| 558 | |
| 559 | =cut |
| 560 | |
| 561 | Note that these macros are repeated in Devel::PPPort, so should also be |
| 562 | patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc |
| 563 | |
| 564 | */ |
| 565 | |
| 566 | /* Specify the widest unsigned type on the platform. Use U64TYPE because U64 |
| 567 | * is known only in the perl core, and this macro can be called from outside |
| 568 | * that */ |
| 569 | #ifdef HAS_QUAD |
| 570 | # define WIDEST_UTYPE U64TYPE |
| 571 | #else |
| 572 | # define WIDEST_UTYPE U32 |
| 573 | #endif |
| 574 | |
| 575 | /* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in |
| 576 | * the lower 8. It is designed to be hopefully bomb-proof, making sure that no |
| 577 | * bits of information are lost even on a 64-bit machine, but to get the |
| 578 | * compiler to optimize it out if possible. This is because Configure makes |
| 579 | * sure that the machine has an 8-bit byte, so if c is stored in a byte, the |
| 580 | * sizeof() guarantees that this evaluates to a constant true at compile time. |
| 581 | */ |
| 582 | #define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF)) |
| 583 | |
| 584 | #ifdef EBCDIC |
| 585 | # define isASCII(c) (FITS_IN_8_BITS(c) && (NATIVE_TO_UNI((U8) (c)) < 128)) |
| 586 | #else |
| 587 | # define isASCII(c) ((WIDEST_UTYPE)(c) < 128) |
| 588 | #endif |
| 589 | |
| 590 | #define isASCII_A(c) isASCII(c) |
| 591 | #define isASCII_L1(c) isASCII(c) |
| 592 | |
| 593 | /* ASCII range only */ |
| 594 | #ifdef H_PERL /* If have access to perl.h, lookup in its table */ |
| 595 | /* Bits for PL_charclass[]. These use names used in l1_char_class_tab.h but |
| 596 | * their actual definitions are here. If that has a name not used here, it |
| 597 | * won't compile. */ |
| 598 | # define _CC_ALNUMC_A (1<<0) |
| 599 | # define _CC_ALNUMC_L1 (1<<1) |
| 600 | # define _CC_ALPHA_A (1<<2) |
| 601 | # define _CC_ALPHA_L1 (1<<3) |
| 602 | # define _CC_BLANK_A (1<<4) |
| 603 | # define _CC_BLANK_L1 (1<<5) |
| 604 | # define _CC_CHARNAME_CONT (1<<6) |
| 605 | # define _CC_CNTRL_A (1<<7) |
| 606 | # define _CC_CNTRL_L1 (1<<8) |
| 607 | # define _CC_DIGIT_A (1<<9) |
| 608 | # define _CC_GRAPH_A (1<<10) |
| 609 | # define _CC_GRAPH_L1 (1<<11) |
| 610 | # define _CC_IDFIRST_A (1<<12) |
| 611 | # define _CC_IDFIRST_L1 (1<<13) |
| 612 | # define _CC_LOWER_A (1<<14) |
| 613 | # define _CC_LOWER_L1 (1<<15) |
| 614 | # define _CC_OCTAL_A (1<<16) |
| 615 | # define _CC_PRINT_A (1<<17) |
| 616 | # define _CC_PRINT_L1 (1<<18) |
| 617 | # define _CC_PSXSPC_A (1<<19) |
| 618 | # define _CC_PSXSPC_L1 (1<<20) |
| 619 | # define _CC_PUNCT_A (1<<21) |
| 620 | # define _CC_PUNCT_L1 (1<<22) |
| 621 | # define _CC_SPACE_A (1<<23) |
| 622 | # define _CC_SPACE_L1 (1<<24) |
| 623 | # define _CC_UPPER_A (1<<25) |
| 624 | # define _CC_UPPER_L1 (1<<26) |
| 625 | # define _CC_WORDCHAR_A (1<<27) |
| 626 | # define _CC_WORDCHAR_L1 (1<<28) |
| 627 | # define _CC_XDIGIT_A (1<<29) |
| 628 | # define _CC_NONLATIN1_FOLD (1<<30) |
| 629 | /* Unused |
| 630 | * (1<<31) |
| 631 | * If more are needed, can give up some of the above. The first ones to go |
| 632 | * would be those that require just two tests to verify, either there are two |
| 633 | * code points, like BLANK_A, or occupy a single range like OCTAL_A, DIGIT_A, |
| 634 | * UPPER_A, and LOWER_A. |
| 635 | */ |
| 636 | |
| 637 | # ifdef DOINIT |
| 638 | EXTCONST U32 PL_charclass[] = { |
| 639 | # include "l1_char_class_tab.h" |
| 640 | }; |
| 641 | |
| 642 | # else /* ! DOINIT */ |
| 643 | EXTCONST U32 PL_charclass[]; |
| 644 | # endif |
| 645 | |
| 646 | # define isALNUMC_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_ALNUMC_A)) |
| 647 | # define isALPHA_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_ALPHA_A)) |
| 648 | # define isBLANK_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_BLANK_A)) |
| 649 | # define isCNTRL_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_CNTRL_A)) |
| 650 | # define isDIGIT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_DIGIT_A)) |
| 651 | # define isGRAPH_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_GRAPH_A)) |
| 652 | # define isIDFIRST_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_IDFIRST_A)) |
| 653 | # define isLOWER_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_LOWER_A)) |
| 654 | # define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_OCTAL_A)) |
| 655 | # define isPRINT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PRINT_A)) |
| 656 | # define isPSXSPC_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PSXSPC_A)) |
| 657 | # define isPUNCT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PUNCT_A)) |
| 658 | # define isSPACE_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_SPACE_A)) |
| 659 | # define isUPPER_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_UPPER_A)) |
| 660 | # define isWORDCHAR_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_WORDCHAR_A)) |
| 661 | # define isXDIGIT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_XDIGIT_A)) |
| 662 | /* Either participates in a fold with a character above 255, or is a |
| 663 | * multi-char fold */ |
| 664 | # define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) ((! cBOOL(FITS_IN_8_BITS(c))) || (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_NONLATIN1_FOLD)) |
| 665 | #else /* No perl.h. */ |
| 666 | # define isOCTAL_A(c) ((c) <= '7' && (c) >= '0') |
| 667 | # ifdef EBCDIC |
| 668 | # define isALNUMC_A(c) (isASCII(c) && isALNUMC(c)) |
| 669 | # define isALPHA_A(c) (isASCII(c) && isALPHA(c)) |
| 670 | # define isBLANK_A(c) (isASCII(c) && isBLANK(c)) |
| 671 | # define isCNTRL_A(c) (isASCII(c) && isCNTRL(c)) |
| 672 | # define isDIGIT_A(c) (isASCII(c) && isDIGIT(c)) |
| 673 | # define isGRAPH_A(c) (isASCII(c) && isGRAPH(c)) |
| 674 | # define isIDFIRST_A(c) (isASCII(c) && isIDFIRST(c)) |
| 675 | # define isLOWER_A(c) (isASCII(c) && isLOWER(c)) |
| 676 | # define isPRINT_A(c) (isASCII(c) && isPRINT(c)) |
| 677 | # define isPSXSPC_A(c) (isASCII(c) && isPSXSPC(c)) |
| 678 | # define isPUNCT_A(c) (isASCII(c) && isPUNCT(c)) |
| 679 | # define isSPACE_A(c) (isASCII(c) && isSPACE(c)) |
| 680 | # define isUPPER_A(c) (isASCII(c) && isUPPER(c)) |
| 681 | # define isWORDCHAR_A(c) (isASCII(c) && isWORDCHAR(c)) |
| 682 | # define isXDIGIT_A(c) (isASCII(c) && isXDIGIT(c)) |
| 683 | # else /* ASCII platform, no perl.h */ |
| 684 | # define isALNUMC_A(c) (isALPHA_A(c) || isDIGIT_A(c)) |
| 685 | # define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c)) |
| 686 | # define isBLANK_A(c) ((c) == ' ' || (c) == '\t') |
| 687 | # define isCNTRL_A(c) (FITS_IN_8_BITS(c) && ((U8) (c) < ' ' || (c) == 127)) |
| 688 | # define isDIGIT_A(c) ((c) <= '9' && (c) >= '0') |
| 689 | # define isGRAPH_A(c) (isWORDCHAR_A(c) || isPUNCT_A(c)) |
| 690 | # define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_') |
| 691 | # define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z') |
| 692 | # define isPRINT_A(c) (((c) >= 32 && (c) < 127)) |
| 693 | # define isPSXSPC_A(c) (isSPACE_A(c) || (c) == '\v') |
| 694 | # define isPUNCT_A(c) (((c) >= 33 && (c) <= 47) || ((c) >= 58 && (c) <= 64) || ((c) >= 91 && (c) <= 96) || ((c) >= 123 && (c) <= 126)) |
| 695 | # define isSPACE_A(c) ((c) == ' ' || (c) == '\t' || (c) == '\n' || (c) =='\r' || (c) == '\f') |
| 696 | # define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A') |
| 697 | # define isWORDCHAR_A(c) (isALPHA_A(c) || isDIGIT_A(c) || (c) == '_') |
| 698 | # define isXDIGIT_A(c) (isDIGIT_A(c) || ((c) >= 'a' && (c) <= 'f') || ((c) <= 'F' && (c) >= 'A')) |
| 699 | # endif |
| 700 | #endif /* ASCII range definitions */ |
| 701 | |
| 702 | /* Latin1 definitions */ |
| 703 | #ifdef H_PERL |
| 704 | # define isALNUMC_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_ALNUMC_L1)) |
| 705 | # define isALPHA_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_ALPHA_L1)) |
| 706 | # define isBLANK_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_BLANK_L1)) |
| 707 | /* continuation character for legal NAME in \N{NAME} */ |
| 708 | # define isCHARNAME_CONT(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_CHARNAME_CONT)) |
| 709 | # define isCNTRL_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_CNTRL_L1)) |
| 710 | # define isGRAPH_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_GRAPH_L1)) |
| 711 | # define isIDFIRST_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_IDFIRST_L1)) |
| 712 | # define isLOWER_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_LOWER_L1)) |
| 713 | # define isPRINT_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PRINT_L1)) |
| 714 | # define isPSXSPC_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PSXSPC_L1)) |
| 715 | # define isPUNCT_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PUNCT_L1)) |
| 716 | # define isSPACE_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_SPACE_L1)) |
| 717 | # define isUPPER_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_UPPER_L1)) |
| 718 | # define isWORDCHAR_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_WORDCHAR_L1)) |
| 719 | #else /* No access to perl.h. Only a few provided here, just in case needed |
| 720 | * for backwards compatibility */ |
| 721 | /* ALPHAU includes Unicode semantics for latin1 characters. It has an extra |
| 722 | * >= AA test to speed up ASCII-only tests at the expense of the others */ |
| 723 | # define isALPHA_L1(c) (isALPHA(c) || (NATIVE_TO_UNI((U8) c) >= 0xAA \ |
| 724 | && ((NATIVE_TO_UNI((U8) c) >= 0xC0 \ |
| 725 | && NATIVE_TO_UNI((U8) c) != 0xD7 && NATIVE_TO_UNI((U8) c) != 0xF7) \ |
| 726 | || NATIVE_TO_UNI((U8) c) == 0xAA \ |
| 727 | || NATIVE_TO_UNI((U8) c) == 0xB5 \ |
| 728 | || NATIVE_TO_UNI((U8) c) == 0xBA))) |
| 729 | # define isCHARNAME_CONT(c) (isALNUM_L1(c) || (c) == ' ' || (c) == '-' || (c) == '(' || (c) == ')' || (c) == ':' || NATIVE_TO_UNI((U8) c) == 0xA0) |
| 730 | #endif |
| 731 | |
| 732 | /* Macros for backwards compatibility and for completeness when the ASCII and |
| 733 | * Latin1 values are identical */ |
| 734 | #define isALNUM(c) isWORDCHAR(c) |
| 735 | #define isALNUMU(c) isWORDCHAR_L1(c) |
| 736 | #define isALPHAU(c) isALPHA_L1(c) |
| 737 | #define isDIGIT_L1(c) isDIGIT_A(c) |
| 738 | #define isOCTAL(c) isOCTAL_A(c) |
| 739 | #define isOCTAL_L1(c) isOCTAL_A(c) |
| 740 | #define isXDIGIT_L1(c) isXDIGIT_A(c) |
| 741 | |
| 742 | /* Macros that differ between EBCDIC and ASCII. Where C89 defines a function, |
| 743 | * that is used in the EBCDIC form, because in EBCDIC we do not do locales: |
| 744 | * therefore can use native functions. For those where C89 doesn't define a |
| 745 | * function, use our function, assuming that the EBCDIC code page is isomorphic |
| 746 | * with Latin1, which the three currently recognized by Perl are. Some libc's |
| 747 | * have an isblank(), but it's not guaranteed. */ |
| 748 | #ifdef EBCDIC |
| 749 | # define isALNUMC(c) isalnum(c) |
| 750 | # define isALPHA(c) isalpha(c) |
| 751 | # define isBLANK(c) ((c) == ' ' || (c) == '\t' || NATIVE_TO_UNI(c) == 0xA0) |
| 752 | # define isCNTRL(c) iscntrl(c) |
| 753 | # define isDIGIT(c) isdigit(c) |
| 754 | # define isGRAPH(c) isgraph(c) |
| 755 | # define isIDFIRST(c) (isALPHA(c) || (c) == '_') |
| 756 | # define isLOWER(c) islower(c) |
| 757 | # define isPRINT(c) isprint(c) |
| 758 | # define isPSXSPC(c) isspace(c) |
| 759 | # define isPUNCT(c) ispunct(c) |
| 760 | # define isSPACE(c) (isPSXSPC(c) && (c) != '\v') |
| 761 | # define isUPPER(c) isupper(c) |
| 762 | # define isXDIGIT(c) isxdigit(c) |
| 763 | # define isWORDCHAR(c) (isalnum(c) || (c) == '_') |
| 764 | # define toLOWER(c) tolower(c) |
| 765 | # define toUPPER(c) toupper(c) |
| 766 | #else /* Not EBCDIC: ASCII-only matching */ |
| 767 | # define isALNUMC(c) isALNUMC_A(c) |
| 768 | # define isALPHA(c) isALPHA_A(c) |
| 769 | # define isBLANK(c) isBLANK_A(c) |
| 770 | # define isCNTRL(c) isCNTRL_A(c) |
| 771 | # define isDIGIT(c) isDIGIT_A(c) |
| 772 | # define isGRAPH(c) isGRAPH_A(c) |
| 773 | # define isIDFIRST(c) isIDFIRST_A(c) |
| 774 | # define isLOWER(c) isLOWER_A(c) |
| 775 | # define isPRINT(c) isPRINT_A(c) |
| 776 | # define isPSXSPC(c) isPSXSPC_A(c) |
| 777 | # define isPUNCT(c) isPUNCT_A(c) |
| 778 | # define isSPACE(c) isSPACE_A(c) |
| 779 | # define isUPPER(c) isUPPER_A(c) |
| 780 | # define isWORDCHAR(c) isWORDCHAR_A(c) |
| 781 | # define isXDIGIT(c) isXDIGIT_A(c) |
| 782 | |
| 783 | /* ASCII casing. These could also be written as |
| 784 | #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c)) |
| 785 | #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c)) |
| 786 | which uses table lookup and mask instead of subtraction. (This would |
| 787 | work because the _MOD does not apply in the ASCII range) */ |
| 788 | # define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c)) |
| 789 | # define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c)) |
| 790 | #endif |
| 791 | |
| 792 | |
| 793 | /* Use table lookup for speed; return error character for input |
| 794 | * out-of-range */ |
| 795 | #define toLOWER_LATIN1(c) (FITS_IN_8_BITS(c) \ |
| 796 | ? UNI_TO_NATIVE(PL_latin1_lc[ \ |
| 797 | NATIVE_TO_UNI( (U8) (c)) ]) \ |
| 798 | : UNICODE_REPLACEMENT) |
| 799 | /* Modified uc. Is correct uc except for three non-ascii chars which are |
| 800 | * all mapped to one of them, and these need special handling; error |
| 801 | * character for input out-of-range */ |
| 802 | #define toUPPER_LATIN1_MOD(c) (FITS_IN_8_BITS(c) \ |
| 803 | ? UNI_TO_NATIVE(PL_mod_latin1_uc[ \ |
| 804 | NATIVE_TO_UNI( (U8) (c)) ]) \ |
| 805 | : UNICODE_REPLACEMENT) |
| 806 | |
| 807 | #ifdef USE_NEXT_CTYPE |
| 808 | |
| 809 | # define isALNUM_LC(c) \ |
| 810 | (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_') |
| 811 | # define isIDFIRST_LC(c) \ |
| 812 | (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_') |
| 813 | # define isALPHA_LC(c) NXIsAlpha((unsigned int)(c)) |
| 814 | # define isSPACE_LC(c) NXIsSpace((unsigned int)(c)) |
| 815 | # define isDIGIT_LC(c) NXIsDigit((unsigned int)(c)) |
| 816 | # define isUPPER_LC(c) NXIsUpper((unsigned int)(c)) |
| 817 | # define isLOWER_LC(c) NXIsLower((unsigned int)(c)) |
| 818 | # define isALNUMC_LC(c) NXIsAlNum((unsigned int)(c)) |
| 819 | # define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c)) |
| 820 | # define isGRAPH_LC(c) NXIsGraph((unsigned int)(c)) |
| 821 | # define isPRINT_LC(c) NXIsPrint((unsigned int)(c)) |
| 822 | # define isPUNCT_LC(c) NXIsPunct((unsigned int)(c)) |
| 823 | # define toUPPER_LC(c) NXToUpper((unsigned int)(c)) |
| 824 | # define toLOWER_LC(c) NXToLower((unsigned int)(c)) |
| 825 | |
| 826 | #else /* !USE_NEXT_CTYPE */ |
| 827 | |
| 828 | # if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII)) |
| 829 | |
| 830 | # define isALNUM_LC(c) (isalnum((unsigned char)(c)) || (char)(c) == '_') |
| 831 | # define isIDFIRST_LC(c) (isalpha((unsigned char)(c)) || (char)(c) == '_') |
| 832 | # define isALPHA_LC(c) isalpha((unsigned char)(c)) |
| 833 | # define isSPACE_LC(c) isspace((unsigned char)(c)) |
| 834 | # define isDIGIT_LC(c) isdigit((unsigned char)(c)) |
| 835 | # define isUPPER_LC(c) isupper((unsigned char)(c)) |
| 836 | # define isLOWER_LC(c) islower((unsigned char)(c)) |
| 837 | # define isALNUMC_LC(c) isalnum((unsigned char)(c)) |
| 838 | # define isCNTRL_LC(c) iscntrl((unsigned char)(c)) |
| 839 | # define isGRAPH_LC(c) isgraph((unsigned char)(c)) |
| 840 | # define isPRINT_LC(c) isprint((unsigned char)(c)) |
| 841 | # define isPUNCT_LC(c) ispunct((unsigned char)(c)) |
| 842 | # define toUPPER_LC(c) toupper((unsigned char)(c)) |
| 843 | # define toLOWER_LC(c) tolower((unsigned char)(c)) |
| 844 | |
| 845 | # else |
| 846 | |
| 847 | # define isALNUM_LC(c) (isascii(c) && (isalnum(c) || (c) == '_')) |
| 848 | # define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_')) |
| 849 | # define isALPHA_LC(c) (isascii(c) && isalpha(c)) |
| 850 | # define isSPACE_LC(c) (isascii(c) && isspace(c)) |
| 851 | # define isDIGIT_LC(c) (isascii(c) && isdigit(c)) |
| 852 | # define isUPPER_LC(c) (isascii(c) && isupper(c)) |
| 853 | # define isLOWER_LC(c) (isascii(c) && islower(c)) |
| 854 | # define isALNUMC_LC(c) (isascii(c) && isalnum(c)) |
| 855 | # define isCNTRL_LC(c) (isascii(c) && iscntrl(c)) |
| 856 | # define isGRAPH_LC(c) (isascii(c) && isgraph(c)) |
| 857 | # define isPRINT_LC(c) (isascii(c) && isprint(c)) |
| 858 | # define isPUNCT_LC(c) (isascii(c) && ispunct(c)) |
| 859 | # define toUPPER_LC(c) toupper(c) |
| 860 | # define toLOWER_LC(c) tolower(c) |
| 861 | |
| 862 | # endif |
| 863 | #endif /* USE_NEXT_CTYPE */ |
| 864 | |
| 865 | #define isPSXSPC_LC(c) (isSPACE_LC(c) || (c) == '\v') |
| 866 | #define isBLANK_LC(c) isBLANK(c) /* could be wrong */ |
| 867 | |
| 868 | /* For use in the macros just below. If the input is Latin1, use the Latin1 |
| 869 | * (_L1) version of the macro; otherwise use the function. Won't compile if |
| 870 | * 'c' isn't unsigned, as won't match function prototype. The macros do bounds |
| 871 | * checking, so have duplicate checks here, so could create versions of the |
| 872 | * macros that don't, but experiments show that gcc optimizes them out anyway. |
| 873 | */ |
| 874 | #define generic_uni(macro, function, c) ((c) < 256 \ |
| 875 | ? CAT2(macro, _L1)(c) \ |
| 876 | : function(c)) |
| 877 | /* Note that all ignore 'use bytes' */ |
| 878 | |
| 879 | #define isALNUM_uni(c) generic_uni(isWORDCHAR, is_uni_alnum, c) |
| 880 | #define isIDFIRST_uni(c) generic_uni(isIDFIRST, is_uni_idfirst, c) |
| 881 | #define isALPHA_uni(c) generic_uni(isALPHA, is_uni_alpha, c) |
| 882 | #define isSPACE_uni(c) generic_uni(isSPACE, is_uni_space, c) |
| 883 | #define isDIGIT_uni(c) generic_uni(isDIGIT, is_uni_digit, c) |
| 884 | #define isUPPER_uni(c) generic_uni(isUPPER, is_uni_upper, c) |
| 885 | #define isLOWER_uni(c) generic_uni(isLOWER, is_uni_lower, c) |
| 886 | #define isASCII_uni(c) isASCII(c) |
| 887 | /* All controls are in Latin1 */ |
| 888 | #define isCNTRL_uni(c) ((c) < 256 && isCNTRL_L1(c)) |
| 889 | #define isGRAPH_uni(c) generic_uni(isGRAPH, is_uni_graph, c) |
| 890 | #define isPRINT_uni(c) generic_uni(isPRINT, is_uni_print, c) |
| 891 | #define isPUNCT_uni(c) generic_uni(isPUNCT, is_uni_punct, c) |
| 892 | #define isXDIGIT_uni(c) generic_uni(isXDIGIT, is_uni_xdigit, c) |
| 893 | #define toUPPER_uni(c,s,l) to_uni_upper(c,s,l) |
| 894 | #define toTITLE_uni(c,s,l) to_uni_title(c,s,l) |
| 895 | #define toLOWER_uni(c,s,l) to_uni_lower(c,s,l) |
| 896 | #define toFOLD_uni(c,s,l) to_uni_fold(c,s,l) |
| 897 | |
| 898 | /* Posix and regular space differ only in U+000B, which is in Latin1 */ |
| 899 | #define isPSXSPC_uni(c) ((c) < 256 ? isPSXSPC_L1(c) : isSPACE_uni(c)) |
| 900 | #define isBLANK_uni(c) isBLANK(c) /* could be wrong */ |
| 901 | |
| 902 | #define isALNUM_LC_uvchr(c) (c < 256 ? isALNUM_LC(c) : is_uni_alnum_lc(c)) |
| 903 | #define isIDFIRST_LC_uvchr(c) (c < 256 ? isIDFIRST_LC(c) : is_uni_idfirst_lc(c)) |
| 904 | #define isALPHA_LC_uvchr(c) (c < 256 ? isALPHA_LC(c) : is_uni_alpha_lc(c)) |
| 905 | #define isSPACE_LC_uvchr(c) (c < 256 ? isSPACE_LC(c) : is_uni_space_lc(c)) |
| 906 | #define isDIGIT_LC_uvchr(c) (c < 256 ? isDIGIT_LC(c) : is_uni_digit_lc(c)) |
| 907 | #define isUPPER_LC_uvchr(c) (c < 256 ? isUPPER_LC(c) : is_uni_upper_lc(c)) |
| 908 | #define isLOWER_LC_uvchr(c) (c < 256 ? isLOWER_LC(c) : is_uni_lower_lc(c)) |
| 909 | #define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : is_uni_cntrl_lc(c)) |
| 910 | #define isGRAPH_LC_uvchr(c) (c < 256 ? isGRAPH_LC(c) : is_uni_graph_lc(c)) |
| 911 | #define isPRINT_LC_uvchr(c) (c < 256 ? isPRINT_LC(c) : is_uni_print_lc(c)) |
| 912 | #define isPUNCT_LC_uvchr(c) (c < 256 ? isPUNCT_LC(c) : is_uni_punct_lc(c)) |
| 913 | |
| 914 | #define isPSXSPC_LC_uni(c) (isSPACE_LC_uni(c) ||(c) == '\f') |
| 915 | #define isBLANK_LC_uni(c) isBLANK(c) /* could be wrong */ |
| 916 | |
| 917 | /* For use in the macros just below. If the input is ASCII, use the ASCII (_A) |
| 918 | * version of the macro; if the input is in the upper Latin1 range, use the |
| 919 | * Latin1 (_L1) version of the macro, after converting from utf8; otherwise use |
| 920 | * the function. This relies on the fact that ASCII characters have the same |
| 921 | * representation whether utf8 or not */ |
| 922 | #define generic_utf8(macro, function, p) (isASCII(*(p)) \ |
| 923 | ? CAT2(CAT2(macro,_),A)(*(p)) \ |
| 924 | : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \ |
| 925 | ? CAT2(macro, _L1) \ |
| 926 | (TWO_BYTE_UTF8_TO_UNI(*(p), \ |
| 927 | *((p)+1))) \ |
| 928 | : function(p)) |
| 929 | |
| 930 | /* Note that all ignore 'use bytes' */ |
| 931 | |
| 932 | #define isALNUM_utf8(p) generic_utf8(isWORDCHAR, is_utf8_alnum, p) |
| 933 | /* To prevent S_scan_word in toke.c from hanging, we have to make sure that |
| 934 | * IDFIRST is an alnum. See |
| 935 | * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you |
| 936 | * ever wanted to know about. XXX It is unclear if this should extend to |
| 937 | * isIDFIRST_uni() which it hasn't so far. (In the ASCII range, there isn't a |
| 938 | * difference.) This used to be not the XID version, but we decided to go with |
| 939 | * the more modern Unicode definition */ |
| 940 | #define isIDFIRST_utf8(p) (isASCII(*(p)) \ |
| 941 | ? isIDFIRST_A(*(p)) \ |
| 942 | : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \ |
| 943 | ? isIDFIRST_L1(TWO_BYTE_UTF8_TO_UNI(*(p), \ |
| 944 | *((p)+1)))\ |
| 945 | : Perl__is_utf8__perl_idstart(aTHX_ p)) |
| 946 | #define isIDCONT_utf8(p) generic_utf8(isWORDCHAR, is_utf8_xidcont, p) |
| 947 | #define isALPHA_utf8(p) generic_utf8(isALPHA, is_utf8_alpha, p) |
| 948 | #define isSPACE_utf8(p) generic_utf8(isSPACE, is_utf8_space, p) |
| 949 | #define isDIGIT_utf8(p) generic_utf8(isDIGIT, is_utf8_digit, p) |
| 950 | #define isUPPER_utf8(p) generic_utf8(isUPPER, is_utf8_upper, p) |
| 951 | #define isLOWER_utf8(p) generic_utf8(isLOWER, is_utf8_lower, p) |
| 952 | /* Because ASCII is invariant under utf8, the non-utf8 macro works */ |
| 953 | #define isASCII_utf8(p) isASCII(p) |
| 954 | #define isCNTRL_utf8(p) generic_utf8(isCNTRL, is_utf8_cntrl, p) |
| 955 | #define isGRAPH_utf8(p) generic_utf8(isGRAPH, is_utf8_graph, p) |
| 956 | #define isPRINT_utf8(p) generic_utf8(isPRINT, is_utf8_print, p) |
| 957 | #define isPUNCT_utf8(p) generic_utf8(isPUNCT, is_utf8_punct, p) |
| 958 | #define isXDIGIT_utf8(p) generic_utf8(isXDIGIT, is_utf8_xdigit, p) |
| 959 | #define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l) |
| 960 | #define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l) |
| 961 | #define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l) |
| 962 | |
| 963 | /* Posix and regular space differ only in U+000B, which is in ASCII (and hence |
| 964 | * Latin1 */ |
| 965 | #define isPSXSPC_utf8(p) ((isASCII(*(p))) \ |
| 966 | ? isPSXSPC_A(*(p)) \ |
| 967 | : (UTF8_IS_DOWNGRADEABLE_START(*(p)) \ |
| 968 | ? isPSXSPC_L1(TWO_BYTE_UTF8_TO_UNI(*(p), \ |
| 969 | *((p)+1)))\ |
| 970 | : isSPACE_utf8(p))) |
| 971 | #define isBLANK_utf8(c) isBLANK(c) /* could be wrong */ |
| 972 | |
| 973 | #define isALNUM_LC_utf8(p) isALNUM_LC_uvchr(utf8_to_uvchr(p, 0)) |
| 974 | #define isIDFIRST_LC_utf8(p) isIDFIRST_LC_uvchr(utf8_to_uvchr(p, 0)) |
| 975 | #define isALPHA_LC_utf8(p) isALPHA_LC_uvchr(utf8_to_uvchr(p, 0)) |
| 976 | #define isSPACE_LC_utf8(p) isSPACE_LC_uvchr(utf8_to_uvchr(p, 0)) |
| 977 | #define isDIGIT_LC_utf8(p) isDIGIT_LC_uvchr(utf8_to_uvchr(p, 0)) |
| 978 | #define isUPPER_LC_utf8(p) isUPPER_LC_uvchr(utf8_to_uvchr(p, 0)) |
| 979 | #define isLOWER_LC_utf8(p) isLOWER_LC_uvchr(utf8_to_uvchr(p, 0)) |
| 980 | #define isALNUMC_LC_utf8(p) isALNUMC_LC_uvchr(utf8_to_uvchr(p, 0)) |
| 981 | #define isCNTRL_LC_utf8(p) isCNTRL_LC_uvchr(utf8_to_uvchr(p, 0)) |
| 982 | #define isGRAPH_LC_utf8(p) isGRAPH_LC_uvchr(utf8_to_uvchr(p, 0)) |
| 983 | #define isPRINT_LC_utf8(p) isPRINT_LC_uvchr(utf8_to_uvchr(p, 0)) |
| 984 | #define isPUNCT_LC_utf8(p) isPUNCT_LC_uvchr(utf8_to_uvchr(p, 0)) |
| 985 | |
| 986 | #define isPSXSPC_LC_utf8(c) (isSPACE_LC_utf8(c) ||(c) == '\f') |
| 987 | #define isBLANK_LC_utf8(c) isBLANK(c) /* could be wrong */ |
| 988 | |
| 989 | /* This conversion works both ways, strangely enough. On EBCDIC platforms, |
| 990 | * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII */ |
| 991 | # define toCTRL(c) (toUPPER(NATIVE_TO_UNI(c)) ^ 64) |
| 992 | |
| 993 | /* Line numbers are unsigned, 32 bits. */ |
| 994 | typedef U32 line_t; |
| 995 | #define NOLINE ((line_t) 4294967295UL) |
| 996 | |
| 997 | /* Helpful alias for version prescan */ |
| 998 | #define is_LAX_VERSION(a,b) \ |
| 999 | (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL)) |
| 1000 | |
| 1001 | #define is_STRICT_VERSION(a,b) \ |
| 1002 | (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL)) |
| 1003 | |
| 1004 | #define BADVERSION(a,b,c) \ |
| 1005 | if (b) { \ |
| 1006 | *b = c; \ |
| 1007 | } \ |
| 1008 | return a; |
| 1009 | |
| 1010 | /* |
| 1011 | =head1 Memory Management |
| 1012 | |
| 1013 | =for apidoc Am|void|Newx|void* ptr|int nitems|type |
| 1014 | The XSUB-writer's interface to the C C<malloc> function. |
| 1015 | |
| 1016 | In 5.9.3, Newx() and friends replace the older New() API, and drops |
| 1017 | the first parameter, I<x>, a debug aid which allowed callers to identify |
| 1018 | themselves. This aid has been superseded by a new build option, |
| 1019 | PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still |
| 1020 | there for use in XS modules supporting older perls. |
| 1021 | |
| 1022 | =for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast |
| 1023 | The XSUB-writer's interface to the C C<malloc> function, with |
| 1024 | cast. See also C<Newx>. |
| 1025 | |
| 1026 | =for apidoc Am|void|Newxz|void* ptr|int nitems|type |
| 1027 | The XSUB-writer's interface to the C C<malloc> function. The allocated |
| 1028 | memory is zeroed with C<memzero>. See also C<Newx>. |
| 1029 | |
| 1030 | =for apidoc Am|void|Renew|void* ptr|int nitems|type |
| 1031 | The XSUB-writer's interface to the C C<realloc> function. |
| 1032 | |
| 1033 | =for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast |
| 1034 | The XSUB-writer's interface to the C C<realloc> function, with |
| 1035 | cast. |
| 1036 | |
| 1037 | =for apidoc Am|void|Safefree|void* ptr |
| 1038 | The XSUB-writer's interface to the C C<free> function. |
| 1039 | |
| 1040 | =for apidoc Am|void|Move|void* src|void* dest|int nitems|type |
| 1041 | The XSUB-writer's interface to the C C<memmove> function. The C<src> is the |
| 1042 | source, C<dest> is the destination, C<nitems> is the number of items, and C<type> is |
| 1043 | the type. Can do overlapping moves. See also C<Copy>. |
| 1044 | |
| 1045 | =for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type |
| 1046 | Like C<Move> but returns dest. Useful for encouraging compilers to tail-call |
| 1047 | optimise. |
| 1048 | |
| 1049 | =for apidoc Am|void|Copy|void* src|void* dest|int nitems|type |
| 1050 | The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the |
| 1051 | source, C<dest> is the destination, C<nitems> is the number of items, and C<type> is |
| 1052 | the type. May fail on overlapping copies. See also C<Move>. |
| 1053 | |
| 1054 | =for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type |
| 1055 | |
| 1056 | Like C<Copy> but returns dest. Useful for encouraging compilers to tail-call |
| 1057 | optimise. |
| 1058 | |
| 1059 | =for apidoc Am|void|Zero|void* dest|int nitems|type |
| 1060 | |
| 1061 | The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the |
| 1062 | destination, C<nitems> is the number of items, and C<type> is the type. |
| 1063 | |
| 1064 | =for apidoc Am|void *|ZeroD|void* dest|int nitems|type |
| 1065 | |
| 1066 | Like C<Zero> but returns dest. Useful for encouraging compilers to tail-call |
| 1067 | optimise. |
| 1068 | |
| 1069 | =for apidoc Am|void|StructCopy|type src|type dest|type |
| 1070 | This is an architecture-independent macro to copy one structure to another. |
| 1071 | |
| 1072 | =for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte |
| 1073 | |
| 1074 | Fill up memory with a byte pattern (a byte repeated over and over |
| 1075 | again) that hopefully catches attempts to access uninitialized memory. |
| 1076 | |
| 1077 | =for apidoc Am|void|PoisonNew|void* dest|int nitems|type |
| 1078 | |
| 1079 | PoisonWith(0xAB) for catching access to allocated but uninitialized memory. |
| 1080 | |
| 1081 | =for apidoc Am|void|PoisonFree|void* dest|int nitems|type |
| 1082 | |
| 1083 | PoisonWith(0xEF) for catching access to freed memory. |
| 1084 | |
| 1085 | =for apidoc Am|void|Poison|void* dest|int nitems|type |
| 1086 | |
| 1087 | PoisonWith(0xEF) for catching access to freed memory. |
| 1088 | |
| 1089 | =cut */ |
| 1090 | |
| 1091 | /* Maintained for backwards-compatibility only. Use newSV() instead. */ |
| 1092 | #ifndef PERL_CORE |
| 1093 | #define NEWSV(x,len) newSV(len) |
| 1094 | #endif |
| 1095 | |
| 1096 | #define MEM_SIZE_MAX ((MEM_SIZE)~0) |
| 1097 | |
| 1098 | /* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil |
| 1099 | * overly eager compilers that will bleat about e.g. |
| 1100 | * (U16)n > (size_t)~0/sizeof(U16) always being false. */ |
| 1101 | #ifdef PERL_MALLOC_WRAP |
| 1102 | #define MEM_WRAP_CHECK(n,t) MEM_WRAP_CHECK_1(n,t,PL_memory_wrap) |
| 1103 | #define MEM_WRAP_CHECK_1(n,t,a) \ |
| 1104 | (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0)) |
| 1105 | #define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t), |
| 1106 | |
| 1107 | #define PERL_STRLEN_ROUNDUP(n) ((void)(((n) > MEM_SIZE_MAX - 2 * PERL_STRLEN_ROUNDUP_QUANTUM) ? (Perl_croak_nocontext("%s",PL_memory_wrap),0):0),((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1))) |
| 1108 | |
| 1109 | #else |
| 1110 | |
| 1111 | #define MEM_WRAP_CHECK(n,t) |
| 1112 | #define MEM_WRAP_CHECK_1(n,t,a) |
| 1113 | #define MEM_WRAP_CHECK_2(n,t,a,b) |
| 1114 | #define MEM_WRAP_CHECK_(n,t) |
| 1115 | |
| 1116 | #define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1))) |
| 1117 | |
| 1118 | #endif |
| 1119 | |
| 1120 | #ifdef PERL_MEM_LOG |
| 1121 | /* |
| 1122 | * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s |
| 1123 | * go through functions, which are handy for debugging breakpoints, but |
| 1124 | * which more importantly get the immediate calling environment (file and |
| 1125 | * line number, and C function name if available) passed in. This info can |
| 1126 | * then be used for logging the calls, for which one gets a sample |
| 1127 | * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined. |
| 1128 | * |
| 1129 | * Known problems: |
| 1130 | * - not all memory allocs get logged, only those |
| 1131 | * that go through Newx() and derivatives (while all |
| 1132 | * Safefrees do get logged) |
| 1133 | * - __FILE__ and __LINE__ do not work everywhere |
| 1134 | * - __func__ or __FUNCTION__ even less so |
| 1135 | * - I think more goes on after the perlio frees but |
| 1136 | * the thing is that STDERR gets closed (as do all |
| 1137 | * the file descriptors) |
| 1138 | * - no deeper calling stack than the caller of the Newx() |
| 1139 | * or the kind, but do I look like a C reflection/introspection |
| 1140 | * utility to you? |
| 1141 | * - the function prototypes for the logging functions |
| 1142 | * probably should maybe be somewhere else than handy.h |
| 1143 | * - one could consider inlining (macrofying) the logging |
| 1144 | * for speed, but I am too lazy |
| 1145 | * - one could imagine recording the allocations in a hash, |
| 1146 | * (keyed by the allocation address?), and maintain that |
| 1147 | * through reallocs and frees, but how to do that without |
| 1148 | * any News() happening...? |
| 1149 | * - lots of -Ddefines to get useful/controllable output |
| 1150 | * - lots of ENV reads |
| 1151 | */ |
| 1152 | |
| 1153 | PERL_EXPORT_C Malloc_t Perl_mem_log_alloc(const UV n, const UV typesize, const char *type_name, Malloc_t newalloc, const char *filename, const int linenumber, const char *funcname); |
| 1154 | |
| 1155 | PERL_EXPORT_C Malloc_t Perl_mem_log_realloc(const UV n, const UV typesize, const char *type_name, Malloc_t oldalloc, Malloc_t newalloc, const char *filename, const int linenumber, const char *funcname); |
| 1156 | |
| 1157 | PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname); |
| 1158 | |
| 1159 | # ifdef PERL_CORE |
| 1160 | # ifndef PERL_MEM_LOG_NOIMPL |
| 1161 | enum mem_log_type { |
| 1162 | MLT_ALLOC, |
| 1163 | MLT_REALLOC, |
| 1164 | MLT_FREE, |
| 1165 | MLT_NEW_SV, |
| 1166 | MLT_DEL_SV |
| 1167 | }; |
| 1168 | # endif |
| 1169 | # if defined(PERL_IN_SV_C) /* those are only used in sv.c */ |
| 1170 | void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname); |
| 1171 | void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname); |
| 1172 | # endif |
| 1173 | # endif |
| 1174 | |
| 1175 | #endif |
| 1176 | |
| 1177 | #ifdef PERL_MEM_LOG |
| 1178 | #define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__) |
| 1179 | #define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__) |
| 1180 | #define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__) |
| 1181 | #endif |
| 1182 | |
| 1183 | #ifndef MEM_LOG_ALLOC |
| 1184 | #define MEM_LOG_ALLOC(n,t,a) (a) |
| 1185 | #endif |
| 1186 | #ifndef MEM_LOG_REALLOC |
| 1187 | #define MEM_LOG_REALLOC(n,t,v,a) (a) |
| 1188 | #endif |
| 1189 | #ifndef MEM_LOG_FREE |
| 1190 | #define MEM_LOG_FREE(a) (a) |
| 1191 | #endif |
| 1192 | |
| 1193 | #define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t)))))) |
| 1194 | #define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t)))))) |
| 1195 | #define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t))))) |
| 1196 | |
| 1197 | #ifndef PERL_CORE |
| 1198 | /* pre 5.9.x compatibility */ |
| 1199 | #define New(x,v,n,t) Newx(v,n,t) |
| 1200 | #define Newc(x,v,n,t,c) Newxc(v,n,t,c) |
| 1201 | #define Newz(x,v,n,t) Newxz(v,n,t) |
| 1202 | #endif |
| 1203 | |
| 1204 | #define Renew(v,n,t) \ |
| 1205 | (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t)))))) |
| 1206 | #define Renewc(v,n,t,c) \ |
| 1207 | (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t)))))) |
| 1208 | |
| 1209 | #ifdef PERL_POISON |
| 1210 | #define Safefree(d) \ |
| 1211 | ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0) |
| 1212 | #else |
| 1213 | #define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d))) |
| 1214 | #endif |
| 1215 | |
| 1216 | #define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t))) |
| 1217 | #define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t))) |
| 1218 | #define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t))) |
| 1219 | |
| 1220 | #define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t))) |
| 1221 | #define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t))) |
| 1222 | #ifdef HAS_MEMSET |
| 1223 | #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t))) |
| 1224 | #else |
| 1225 | /* Using bzero(), which returns void. */ |
| 1226 | #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d) |
| 1227 | #endif |
| 1228 | |
| 1229 | #define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t))) |
| 1230 | #define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB) |
| 1231 | #define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF) |
| 1232 | #define Poison(d,n,t) PoisonFree(d,n,t) |
| 1233 | |
| 1234 | #ifdef USE_STRUCT_COPY |
| 1235 | #define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s))) |
| 1236 | #else |
| 1237 | #define StructCopy(s,d,t) Copy(s,d,1,t) |
| 1238 | #endif |
| 1239 | |
| 1240 | #define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0])) |
| 1241 | #define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0])) |
| 1242 | |
| 1243 | #ifdef NEED_VA_COPY |
| 1244 | # ifdef va_copy |
| 1245 | # define Perl_va_copy(s, d) va_copy(d, s) |
| 1246 | # else |
| 1247 | # if defined(__va_copy) |
| 1248 | # define Perl_va_copy(s, d) __va_copy(d, s) |
| 1249 | # else |
| 1250 | # define Perl_va_copy(s, d) Copy(s, d, 1, va_list) |
| 1251 | # endif |
| 1252 | # endif |
| 1253 | #endif |
| 1254 | |
| 1255 | /* convenience debug macros */ |
| 1256 | #ifdef USE_ITHREADS |
| 1257 | #define pTHX_FORMAT "Perl interpreter: 0x%p" |
| 1258 | #define pTHX__FORMAT ", Perl interpreter: 0x%p" |
| 1259 | #define pTHX_VALUE_ (void *)my_perl, |
| 1260 | #define pTHX_VALUE (void *)my_perl |
| 1261 | #define pTHX__VALUE_ ,(void *)my_perl, |
| 1262 | #define pTHX__VALUE ,(void *)my_perl |
| 1263 | #else |
| 1264 | #define pTHX_FORMAT |
| 1265 | #define pTHX__FORMAT |
| 1266 | #define pTHX_VALUE_ |
| 1267 | #define pTHX_VALUE |
| 1268 | #define pTHX__VALUE_ |
| 1269 | #define pTHX__VALUE |
| 1270 | #endif /* USE_ITHREADS */ |
| 1271 | |
| 1272 | /* Perl_deprecate was not part of the public API, and did not have a deprecate() |
| 1273 | shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor |
| 1274 | CPAN::Unpack show any users outside the core. */ |
| 1275 | #ifdef PERL_CORE |
| 1276 | # define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated") |
| 1277 | #endif |
| 1278 | |
| 1279 | /* |
| 1280 | * Local variables: |
| 1281 | * c-indentation-style: bsd |
| 1282 | * c-basic-offset: 4 |
| 1283 | * indent-tabs-mode: t |
| 1284 | * End: |
| 1285 | * |
| 1286 | * ex: set ts=8 sts=4 sw=4 noet: |
| 1287 | */ |